KernelFusion: Assumption-Free Blind Super-Resolution via Patch Diffusion

• URL: http://arxiv.org/abs/2503.21907v1
• Date: Thu, 27 Mar 2025 18:37:09 GMT
• Title: KernelFusion: Assumption-Free Blind Super-Resolution via Patch Diffusion
• Authors: Oliver Heinimann, Assaf Shocher, Tal Zimbalist, Michal Irani,
• Abstract summary: We introduce a zero-shot diffusion-based method that makes no assumptions about the kernel.<n>We first train an image-specific patch-based diffusion model on the single LR input image, capturing its unique internal patch statistics.<n>We then reconstruct a larger HR image with the same learned patch distribution, while simultaneously recovering the correct downscaling SR- Kernel.
• Score: 13.468846462250168
• License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/
• Abstract: Traditional super-resolution (SR) methods assume an ``ideal'' downscaling SR-kernel (e.g., bicubic downscaling) between the high-resolution (HR) image and the low-resolution (LR) image. Such methods fail once the LR images are generated differently. Current blind-SR methods aim to remove this assumption, but are still fundamentally restricted to rather simplistic downscaling SR-kernels (e.g., anisotropic Gaussian kernels), and fail on more complex (out of distribution) downscaling degradations. However, using the correct SR-kernel is often more important than using a sophisticated SR algorithm. In ``KernelFusion'' we introduce a zero-shot diffusion-based method that makes no assumptions about the kernel. Our method recovers the unique image-specific SR-kernel directly from the LR input image, while simultaneously recovering its corresponding HR image. KernelFusion exploits the principle that the correct SR-kernel is the one that maximizes patch similarity across different scales of the LR image. We first train an image-specific patch-based diffusion model on the single LR input image, capturing its unique internal patch statistics. We then reconstruct a larger HR image with the same learned patch distribution, while simultaneously recovering the correct downscaling SR-kernel that maintains this cross-scale relation between the HR and LR images. Empirical results show that KernelFusion vastly outperforms all SR baselines on complex downscaling degradations, where existing SotA Blind-SR methods fail miserably. By breaking free from predefined kernel assumptions, KernelFusion pushes Blind-SR into a new assumption-free paradigm, handling downscaling kernels previously thought impossible.

Related papers

• Latent Diffusion, Implicit Amplification: Efficient Continuous-Scale Super-Resolution for Remote Sensing Images [7.920423405957888]
  E$2$DiffSR achieves superior objective metrics and visual quality compared to the state-of-the-art SR methods. It reduces the inference time of diffusion-based SR methods to a level comparable to that of non-diffusion methods.
  arXiv  Detail & Related papers  (2024-10-30T09:14:13Z)
• Burst Super-Resolution with Diffusion Models for Improving Perceptual Quality [12.687175237915019]
  Prior SR networks accepting the burst LR images are trained in a deterministic manner, which is known to produce a blurry SR image. Since such blurry images are perceptually degraded, we aim to reconstruct the sharp high-fidelity boundaries. In our proposed method, on the other hand, burst LR features are used to reconstruct the initial burst SR image.
  arXiv  Detail & Related papers  (2024-03-28T13:58:05Z)
• BlindDiff: Empowering Degradation Modelling in Diffusion Models for Blind Image Super-Resolution [52.47005445345593]
  BlindDiff is a DM-based blind SR method to tackle the blind degradation settings in SISR. BlindDiff seamlessly integrates the MAP-based optimization into DMs. Experiments on both synthetic and real-world datasets show that BlindDiff achieves the state-of-the-art performance.
  arXiv  Detail & Related papers  (2024-03-15T11:21:34Z)
• Kernelized Back-Projection Networks for Blind Super Resolution [14.422263486129811]
  Super Resolution (SR) fails to super-resolve Low-Resolution (LR) images degraded by arbitrary degradations. This paper reveals that non-blind SR that is trained simply with various blur kernels exhibits comparable performance as those with the degradation model for blind SR.
  arXiv  Detail & Related papers  (2023-02-16T18:35:39Z)
• Kernel-aware Raw Burst Blind Super-Resolution [0.0]
  Burst super-resolution (SR) provides a possibility of restoring rich details from low-quality images. Existing non-blind networks usually lead to a severe performance drop in recovering high-resolution (HR) images. We introduce a kernel-aware deformable alignment module which can effectively align the raw images with consideration of the blurry priors.
  arXiv  Detail & Related papers  (2021-12-14T11:49:13Z)
• Spectrum-to-Kernel Translation for Accurate Blind Image Super-Resolution [33.59749785182318]
  We propose a novel blind SR framework to super-resolve LR images degraded by arbitrary blur kernel. We first demonstrate that feature representation in frequency domain is more conducive for blur kernel reconstruction than in spatial domain. Experiments on both synthetic and real-world images demonstrate that our proposed method sufficiently reduces blur kernel estimation error.
  arXiv  Detail & Related papers  (2021-10-23T06:03:22Z)
• Mutual Affine Network for Spatially Variant Kernel Estimation in Blind Image Super-Resolution [130.32026819172256]
  Existing blind image super-resolution (SR) methods mostly assume blur kernels are spatially invariant across the whole image. This paper proposes a mutual affine network (MANet) for spatially variant kernel estimation.
  arXiv  Detail & Related papers  (2021-08-11T16:11:17Z)
• End-to-end Alternating Optimization for Blind Super Resolution [68.395664154041]
  Two-step solution involves two independently trained models, which may not be well compatible with each other. We adopt an alternating optimization algorithm, which can estimate the blur kernel and restore the SR image in a single model. Our model can largely outperform state-of-the-art methods and produce more visually favorable results at a much higher speed.
  arXiv  Detail & Related papers  (2021-05-14T15:05:30Z)
• Frequency Consistent Adaptation for Real World Super Resolution [64.91914552787668]
  We propose a novel Frequency Consistent Adaptation (FCA) that ensures the frequency domain consistency when applying Super-Resolution (SR) methods to the real scene. We estimate degradation kernels from unsupervised images and generate the corresponding Low-Resolution (LR) images. Based on the domain-consistent LR-HR pairs, we train easy-implemented Convolutional Neural Network (CNN) SR models.
  arXiv  Detail & Related papers  (2020-12-18T08:25:39Z)
• Unfolding the Alternating Optimization for Blind Super Resolution [60.39409203244474]
  Two-step solution involves two independently trained models, which may not be well compatible with each other. We adopt an alternating optimization algorithm, which can estimate blur kernel and restore SR image in a single model. Our model can largely outperform state-of-the-art methods and produce more visually favorable results at much higher speed.
  arXiv  Detail & Related papers  (2020-10-06T11:27:27Z)
• Cross-Scale Internal Graph Neural Network for Image Super-Resolution [147.77050877373674]
  Non-local self-similarity in natural images has been well studied as an effective prior in image restoration. For single image super-resolution (SISR), most existing deep non-local methods only exploit similar patches within the same scale of the low-resolution (LR) input image. This is achieved using a novel cross-scale internal graph neural network (IGNN)
  arXiv  Detail & Related papers  (2020-06-30T10:48:40Z)

This list is automatically generated from the titles and abstracts of the papers in this site.

This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.